WO2009113149A1

WO2009113149A1 - Device for measuring pattern length and method for measuring pattern length

Info

Publication number: WO2009113149A1
Application number: PCT/JP2008/054295
Authority: WO
Inventors: 純松本; 祥明小木曽
Original assignee: 株式会社アドバンテスト
Priority date: 2008-03-10
Filing date: 2008-03-10
Publication date: 2009-09-17
Also published as: US8431895B2; JP5066252B2; US20110049362A1; DE112008003774T5; TWI375786B; JPWO2009113149A1; TW200942800A

Abstract

A device for measuring a pattern length measures an edge-to-edge spacing of a pattern with excellent repeatability and accuracy. The device for measuring a pattern length has an electron beam irradiation section to irradiate an electron beam onto a specimen while scanning, an image data acquisition section to acquire an image of the pattern in accordance with the amount of electrons generated on the specimen on which the pattern is formed through irradiation of an electron beam, a measurement target area setting section to set up a pair of measurement areas including edges of the pattern image, and a control section which detects shapes of edges on the pattern within the measurement areas and calculates an edge-to-edge spacing of the pattern within a pair of measurement areas. The control section calculates an edge characteristic curve by taking moving average with a specified moving average width for an edge profile where edges on the pattern within the measurement areas are shown with positional coordinates of measurement points at specified intervals and specifies a position of a peak value on the edge characteristic curve as an edge position on the pattern within the measurement areas.

Description

Pattern length measuring device and pattern length measuring method

The present invention relates to a pattern length measuring apparatus and a length measuring method using an electron beam, and more particularly, to a pattern length measuring apparatus and a pattern length measuring method capable of measuring the distance between pattern edges with good reproducibility.

Measured by a scanning electron microscope as a pattern line width measuring method. In a scanning electron microscope, incident electrons are irradiated while being scanned within an electron beam scanning range, secondary electrons emitted from a sample are acquired through a scintillator, and the amount of electrons acquired is converted into luminance to obtain an image. Data is acquired and displayed on the display device.

When managing the characteristics of a semiconductor device using such a scanning electron microscope, work is performed to determine whether the line width of the line pattern and the distance between the patterns are within the design reference value. It is generally adopted. The management of the line width of patterns and the distance between patterns is performed by the following procedure. After moving on the XY stage to the predetermined position of the pattern formed on the photomask, irradiate the electron beam within the specified range centering on the measurement position, and image of the luminance distribution based on the secondary electrons reflected from the irradiation point To get. Then, the waveform of the luminance distribution is acquired from the luminance information of the measurement point and analyzed to obtain the pattern edge position as the line width. Further, the distance between the edges of the two patterns is calculated from the value of the acquired image data. It is determined whether or not the line width of the pattern and the distance between the patterns are within an allowable error range, which is used as a criterion for determining whether the photomask quality is good or as process feedback information to the previous process.

Thus, the measurement of the line width of patterns and the distance between patterns is important in the photomask manufacturing process, and various techniques for measuring the line width and the like have been proposed.

In general, the position where the gradient of luminance corresponding to the amount of secondary electrons is maximized is set as the edge position of the pattern. However, in Patent Document 1, the edge detection in which the position where the secondary electron signal takes the minimum value is regarded as the edge position. A method is disclosed.

As described above, when measuring the line width of a pattern using a scanning electron microscope, the position where the gradient of luminance is maximum is set as the edge position, or the position where the secondary electron signal takes the minimum value is set as the edge. The position method is adopted.

In edge detection of such a pattern, the edge measurement range is widened and the edge position is determined using a plurality of detection values. As a result, the edge position is detected with good reproducibility even if the area designated as the measurement target is slightly shifted.

However, for example, when detecting the edge position of a narrow portion such as the end portion of the line pattern, the measurement range cannot be widened, and it is difficult to detect a stable edge position. In addition, the region designated as the measurement target is shifted, so that the curved portion of the corner portion of the pattern may or may not be included, and it becomes difficult to detect the edge position with high reproducibility.
Japanese Patent Laid-Open No. 5-296754

The present invention has been made in view of the problems of the prior art, and an object thereof is to stably detect the edge positions of a pattern and to measure the distance between the pattern edges with high reproducibility and accuracy. It is to provide a length measuring device and a pattern length measuring method.

The above-described problems include an electron beam irradiation unit that irradiates an electron beam while scanning the sample, and the pattern based on the amount of electrons generated from the sample on which the pattern is formed by the electron beam irradiation. An image data acquisition unit for acquiring an image, a measurement target region setting unit for setting a pair of measurement regions including an edge of the image of the pattern, and detecting the edge shape of the pattern in the measurement region, and the pair of measurement regions A control unit that calculates a distance between the edges of the pattern in the pattern, and the control unit is a flat part of the edge profile in which the edge of the pattern in the measurement region is indicated by the position coordinates of the measurement points at a predetermined interval And the average position of the flat portion is used as the edge position of the pattern in the measurement region.

In the pattern length measuring apparatus according to this aspect, when the flat portion is not detected, the control unit is a value obtained by subtracting a predetermined value assumed to be curved when a pattern is formed from a design value of the pattern width. The edge profile may be averaged by calculating an edge characteristic curve, and the position of the peak value of the edge characteristic curve may be set as the edge position of the pattern.

Moreover, in the pattern length measuring device according to this aspect, the control unit has a y-coordinate value of the edge position of the pattern in one measurement region of the pair of measurement regions facing in the y-axis direction. The y coordinate value of the edge position of the pattern in the other measurement region is y2, and when y1> y2, y1-y2 may be the distance between the edges in the y-axis direction. Is the x coordinate value of the edge position of the pattern in one measurement region of the pair of measurement regions facing in the x-axis direction, and the x coordinate value of the edge position of the pattern in the other measurement region is When the value is x2 and x1 <x2, x2-x1 may be the distance between the edges in the x-axis direction, and the control unit averages the edge profile with a predetermined number of measurement points. And smoothing the pair To detect the shortest distance between the edge of the pattern in one measurement area and the edge of the pattern in the other measurement area using the position coordinates of each measurement point of the smoothed edge of the pattern in the fixed area It may be.

In the present invention, in the measurement between the edges of the pattern, the flat portion is detected from the edge profile that represents the shape of the edge of the pattern in the designated measurement area by the position coordinates, and the average position is calculated as the edge position. ing. As one method for detecting the flat portion, the edge profile is differentiated, and the position where the differential value is first zero is obtained as the start position of the flat portion and the position where the differential value is finally zero as the end position. Can do. In addition, when a flat portion cannot be detected because there is no point where the differential value becomes zero, the profile of the flat portion estimated from the design value of the pattern width is moved and averaged to obtain an edge characteristic curve (moving average profile). And the peak position is calculated as the edge center position. Since the moving average width when performing the moving average uses the length of the flat portion of the edge of the pattern, the center position of the flat portion of the edge can be uniquely obtained from the calculated edge characteristics. Then, the distance between the edges of the pattern in the pair of designated measurement areas is calculated using the edge position uniquely acquired. Thereby, it is possible to prevent the position of the edge from being changed due to the shift of the measurement region, and it is possible to calculate the distance between the edges with high reproducibility.

According to another aspect of the present invention, there is provided a pattern length measuring method implemented in the pattern length measuring apparatus according to the above aspect. The pattern length measuring method according to the embodiment includes a step of obtaining an image of a pattern of a measurement target region, and an edge profile in which the edge of the pattern in the measurement target region is indicated by position coordinates of measurement points at a predetermined interval. A step of detecting, a step of detecting a flat portion of the edge profile, a step of calculating an average position of the flat portion to determine an edge position of the pattern, and a pair of patterns in the pair of measurement target regions Detecting the distance between the edges.

In the pattern length measurement method according to this aspect, when the flat portion is not detected, the step of detecting the flat portion and the step of determining the edge position are performed when a pattern is formed from a design value of the width of the pattern. Calculating an edge characteristic curve by moving and averaging the edge profile with a value obtained by subtracting a predetermined value assumed to be curved, and determining a peak value position of the edge characteristic curve as an edge position of the pattern You may make it be.

Further, in the pattern length measuring method according to this aspect, the step of detecting the distance between the edges includes the y-coordinate value of the edge position in one of the pair of measurement areas facing in the y-axis direction. Is y1, the y-coordinate value of the edge position in the other measurement region is y2, and when y1> y2, y1-y2 is the distance in the y-axis direction between the edges. In the step of detecting the distance between the edges, the x-coordinate value of the edge position in one measurement region of the pair of measurement regions facing in the x-axis direction is x1, and in the other measurement region When the value of the x coordinate of the edge position is x2 and x1 <x2, the step may be to set x2-x1 as the distance between the edges in the x-axis direction, and the distance between the edges is detected. Step is said edge The average of the profile is smoothed at a predetermined number of measurement points, and each coordinate position of the smoothed edge of the pattern in the pair of measurement areas is used to measure the edge of the pattern in one measurement area and the other measurement. And detecting a shortest distance between the edges of the pattern in the region.

FIG. 1 is a configuration diagram of a scanning electron microscope used in an embodiment of the present invention. 2A to 2D are explanatory diagrams of electronic images and profiles acquired by the signal processing unit. FIGS. 3A to 3C are diagrams for explaining a ROI for designating a pattern to be measured and a measurement target region. FIG. 4 is a diagram for explaining the relationship between the edge shape and the moving average profile. 5A to 5D are diagrams for explaining the detection accuracy of the edge position with respect to the value of the moving average width. FIGS. 6A to 6H are diagrams showing examples of opposing edge shapes. FIG. 7 is a diagram for explaining edge shape determination. FIG. 8 is a flowchart illustrating an example of processing for calculating the distance between edges. FIG. 9 is a flowchart illustrating an example of processing for detecting the edge position of the pattern in the ROI. FIGS. 10A to 10D are diagrams for explaining a method for detecting an edge position in the ROI.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.

First, the configuration of a scanning electron microscope used as a pattern length measuring device will be described. Next, a method for measuring the line width of a general pattern will be described. Next, length measurement between edges of a pattern including detection of edge positions with good reproducibility will be described, and finally, a length measurement method between edges of the pattern will be described.

(Configuration of scanning electron microscope)
FIG. 1 is a configuration diagram of a scanning electron microscope according to the present embodiment.

The scanning electron microscope 100 includes an electronic scanning unit 10, a signal processing unit 30, an image display unit 40, a measurement target region setting unit 50, a storage unit 55, an electronic scanning unit 10, a signal processing unit 30, and an image. The display unit 40, the measurement target region setting unit 50, and the control unit 20 that controls each unit of the storage unit 55 are roughly classified. The control unit 20 includes a profile creation unit 21, a differential profile creation unit 22, an edge detection unit 23, and an inter-edge distance measurement unit 24.

The electron scanning unit 10 includes an electron gun 1, a condenser lens 2, a deflection coil 3, an objective lens 4, a moving stage 5, and a sample holder 6.

The charged particles 9 irradiated from the electron gun 1 are irradiated to the sample 7 on the moving stage 5 through the condenser lens 2, the deflection coil 3 and the objective lens 4.

A charged particle 9 (primary electron beam) is irradiated onto the sample 7 while being scanned two-dimensionally, and secondary electrons emitted from the irradiated part are detected by an electron detector 8 composed of a scintillator or the like. The detected amount of secondary electrons is converted into a digital amount by the AD converter of the signal processing unit 30 and stored in the storage unit 55 as image data. The image data is converted into a luminance signal and displayed on the image display unit 40. The image data is arranged on the two-dimensional array so as to have the same arrangement as the scanning position of the primary electron beam on the sample 7, and a two-dimensional digital image is obtained. Each pixel (pixel) of the two-dimensional digital image represents luminance data with an 8-bit information amount.

The electronic deflection amount of the deflection coil 3 and the image scan amount of the image display unit 40 are controlled by the control unit 20. Further, the control unit 20 stores a program for executing the inter-edge measurement.

The profile creation unit 21 creates a line profile representing the luminance signal of the SEM image data in the specified range. The line profile represents a luminance signal corresponding to the amount of secondary electrons, and is considered to reflect the cross-sectional shape of the measurement pattern.

The differential profile creation unit 22 performs a primary differentiation process on the line profile to create a primary differential profile.

The edge detection unit 23 detects the edge of the pattern from the line profile and the primary differential profile. Further, as will be described later, an edge profile indicating the shape of the edge is created, and the edge profile is moved and averaged by a predetermined moving average width to create a moving average profile. The edge position is detected based on this moving average profile.

The inter-edge distance measuring unit 24 measures the distance between the edges of the patterns included in the pair of ROIs set by the measurement target region setting unit 50. The distance between the edges of the two patterns in the x-axis or y-axis direction and the shortest distance between the edges of the two patterns are measured.

(General line width measurement method)
Next, a general method for measuring the line width of the sample pattern shown in FIG. 2A using the scanning electron microscope 100 shown in FIG. 1 will be described.

As the sample 7, as shown in FIG. 2A, a sample in which a wiring pattern 51 is formed on a photomask substrate 50 is used. A part of the sample 7 has a planar shape as shown in FIG. Here, a portion surrounded by a broken line 52 indicates an observation region of the scanning electron microscope 100.

In FIG. 2B, the electron quantity such as secondary electrons obtained by scanning the electron beam on the sample shown in FIG. 2A is detected by the electron detector 8, and the detected electron quantity is converted into a luminance signal. An example of the SEM image displayed by synchronizing the scanning of the electron beam and the scanning of the CRT of the display device is shown.

2) A SEM image is extracted from the SEM image shown in FIG. The length measurement area is an area having a width H of 400 pixels and a length L, for example. This region is selected by the operator by the upper line marker LM1, the lower line marker LM2, the left line marker LM3, and the right line marker LM4.

The H direction of the length measurement area is divided from the extracted SEM image pixel data, and a line profile corresponding to the luminance distribution is obtained for the divided area. When obtaining a line profile, a noise component can be reduced by performing a smoothing process with a width of, for example, 3 pixels in the length L direction.

FIG. 2C is a diagram showing a line profile corresponding to the amount of secondary electrons emitted from the sample obtained when the electron beam is irradiated along the line II in FIG. is there. As shown in FIG. 2C, the line profile (contrast profile) changes abruptly at the edge of the pattern. In order to obtain a position that changes rapidly, the line profile is differentiated to obtain the maximum peak and the minimum peak of the differential signal amount.

Further, as shown in FIG. 2D, the differential waveforms C1 and C2 are obtained by interpolating between the plurality of differential signals Dx before and after the peak to obtain the first peak P1 and the second peak P2 with a resolution of 1/100. The peak position of is calculated. The width W1 of the line pattern is obtained as the distance between the first peak P1 and the second peak P2.

The above processing is performed in each divided area, and the average value of the pattern widths calculated in each area is used as a length measurement value, whereby a more accurate line pattern width W1 can be obtained.

(Measurement between pattern edges)
In the above-described length measurement between edges of a general pattern, an accurate distance is obtained by taking a wide measurement target region. In the present embodiment, even when the measurement target area cannot be widened, accurate measurement of the distance between the edges is possible.

FIG. 3 is a diagram showing an example of measuring the distance between the edges of the pattern. FIGS. 3A to 3C each show a part of an SEM image of a pattern formed on a sample. If the vertical direction in FIG. 3 is the y-axis direction and the horizontal direction is the x-axis direction, for example, FIG. 3A shows patterns P11 and P12 facing in the y-axis direction, and the measurement target range is the ROI box (ROI11). And the ROI 12), the distance between the edges of these patterns is measured. FIG. 3B shows patterns P21 and P21 that face each other in the y-axis direction and are shifted in the x-axis direction. FIG. 3C shows patterns P31 and P32 facing each other in the x-axis direction.

As shown in FIG. 3A, when measuring the distance between the end portions of two patterns, the flat portion has a short length, so a wide range of data cannot be obtained, and accurate measurement is performed. Is difficult. In addition, when the flat part is short as described above, it is applied to the corner with a slight deviation of the ROI box, and when the corner is curved unlike the design value, it depends on whether or not the ROI box is applied to the curved part. The value of the edge position changes greatly. As a result, the detection distance between the edges is different for each length measurement, and the reproducibility of the distance detection between the edges is deteriorated.

In contrast, the inventor of the present application paid attention to the use of the shape of the edge portion in order to uniquely detect the edge position. That is, it was noted that the range of the flat part or the center position of the flat part can be uniquely obtained by analyzing the flat measurement part and the edge shapes at both ends thereof.

In this embodiment, for example, as shown in FIG. 3A, the measurement target region (ROI11 and ROI12) is specified including the curved portion of the corner portion. It will be described below that the range of the flat part to be measured or the center position of the flat part is uniquely calculated from the edge profile of the pattern in the measurement target region (also referred to as ROI box) set in this way.

FIG. 4 is a view showing an example of the tip portion of the pattern. Also, the broken line L1 in FIG. 4 indicates the design value, and the solid line L2 indicates the edge profile of the formed pattern. As shown in FIG. 4, the actually formed pattern is not formed according to the design value, and in particular, the corner is formed with roundness.

A profile obtained by differentiating the profile of the edge shape of L2 is obtained, and the position where the differential value becomes zero first is the start position of the flat portion, and the position where the differential value finally becomes zero is the end position of the flat portion. The distance between the two points can be determined as the range of the flat portion.

In the present embodiment, the position obtained by averaging the range of the flat portion is used as the edge position. When the flat portion is not detected, the length of the portion of the pattern edge that can be estimated to be flat is set as the moving average width, and the moving average profile is created by moving the edge profile using the moving average width. The length of the portion of the edge of the pattern that can be estimated to be flat is a value obtained by subtracting the length assumed to form a corner when the pattern is formed from the design value of the edge. For example, the length assumed to be formed in a curved shape is 50 nm regardless of the length of the edge.

The edge profile represents the edge shape by indicating the edge in position coordinates for each measurement point at a predetermined interval. A moving average process is performed on each measurement point of the edge profile. That is, when the moving average width is calculated to be, for example, 100 nm, if the number of measurement points of the edge profile corresponding to 100 nm is, for example, 36, averaging is performed at the 36 points before and after each measurement point of the edge profile.

As a result of moving average of the edge profile, as shown in FIG. 4, the moving average profile L3 has a peak value, and the position of the peak value corresponds to the center of the flat portion of the edge. This position is called the center edge position at the edge of the pattern in the measurement target region.

If the moving average width is the optimum value, that is, the width of the flat portion of the edge of the measurement target range, the alignment position and the average position can be calculated simultaneously. However, it is difficult to accurately predict the length of the curved portion of the corner of the pattern, and it is difficult to set the moving average width to an optimum value.

The edge position detection accuracy with respect to the moving average width value will be described with reference to FIG. FIG. 5A shows an example of a pattern whose width in the x direction is W5. The length of the flat portion of the edge of this pattern is the length obtained by subtracting the curved length of the corner portion of the pattern, and the optimum moving average width W is obtained. FIG. 5B shows a moving average profile m1 obtained by moving average with a moving average width W5 larger than the optimum moving average width W with respect to the edge profile E1 of FIG. As shown in FIG. 5B, although the moving average profile m1 has a peak value, the y coordinate value is different from the actual y coordinate value, and a difference d5 occurs. An error occurs in the (x, y) coordinate value (also referred to as an average position) of the center position of the edge, and the accuracy of the average position is deteriorated.

FIG. 5 (c) shows a moving average profile m2 obtained by moving and averaging the edge profile E1 with the optimum moving average width W. In this case, the peak position of the moving average profile m2 indicates the center of the edge.

On the other hand, FIG. 5D shows a case where the edge profile E1 is moving averaged with a moving average width W6 smaller than the optimum moving average width W. In this case, the moving average profile m3 has a flat portion with a width d6 in which the y coordinate value is constant without taking a peak value. Therefore, the x coordinate value (also referred to as alignment position) of the edge cannot be uniquely determined. Therefore, the accuracy of the alignment position is deteriorated.

Even if the moving average width is not the optimum value, the alignment position and the average position can be calculated with high accuracy by separately calculating the alignment position and the average position as follows.

First, the moving average width is set to a value W5 larger than the optimum value W, and the moving average width W5 is used to average the edge profile. The moving average width W5 is, for example, a pattern width. Thereby, the alignment position, that is, the center position CL of the flat portion of the edge is uniquely calculated. *

Next, the average value of the y coordinate of the edge profile E1 between −Wm / 2 and + Wm / 2 (see FIG. 5B) is obtained with respect to the x coordinate specified by the alignment position calculation. Note that Wm is a value that considers the variation in the width of the flat portion in the actual SEM image, and is preferably a value close to the optimum value W. In this way, the coordinate value (x, y) of the average position is calculated.

In the length measurement between the edges described above, the case where the pair of edges in the measurement target region are both convex patterns has been described. The shape of the edge to be measured is not limited to a convex shape, but may be a concave shape or a flat shape. Therefore, detection of the edge position corresponding to the shape of the edge will be described below.

FIG. 6 shows an example of the shape of opposing edges. FIG. 6A shows a case where the end portions of the left and right patterns P61L and P61R are convex, and FIG. 6B shows a case where the end portions of the left and right patterns P62L and P62R are concave. FIGS. 6C and 6D show the case where one of the left and right patterns (P63L, P63R, P64L, P64R) is convex and the other is concave. FIGS. 6E and 6F show the case where one of the left and right patterns (P65L, P65R, P66L, and P66R) is convex and the other is flat, and FIGS. 6G and 6H. Indicates a case where one of the left and right patterns (P67L, P67R, P68L, P68R) is concave and the other is flat.

For example, assuming that the right direction in FIG. 6 is the + x direction, in the pattern P61L on the left side in FIG. 6A, after moving average the edge profile, the peak of the moving average profile becomes convex in the + x direction. In the right side pattern P61R, the peak of the moving average profile is convex in the −x direction. In the left pattern P62L in FIG. 6B, the peak of the moving average profile is convex in the −x direction, and in the right pattern P62R in FIG. 6B, the peak of the moving average profile is convex in the −x direction. Become. The position coordinates corresponding to these peak values are set as the center edge position. The same applies to FIGS. 6C and 6D.

Also, as in the pattern P65R on the right side of FIG. 6E, when the edge is a straight line, a peak value cannot be obtained even if moving average is performed, and the center edge is obtained using the peak value of the moving average profile. The position cannot be detected. In this case, the position coordinates of the center edge position of the opposing convex edge are used. For example, in the case of FIG. 6E, when the center edge position of the left pattern P65L is C65L (x1, y1) and the x coordinate of the right pattern P65R is x2, the center edge of the right straight edge Let position C65R be (x2, y1). Similarly, in the case of FIGS. 6F to 6H, the center edge position is detected.

Judgment of edge shape is performed as follows. FIG. 7 shows a part of the edge E in the measurement target region. As shown in FIG. 7, the edge positions detected by dividing the entire edge E into three parts are respectively A (xa, ya), B (xb, yb), and C (xc, yc). For example, when determining the unevenness of the edge in the x-axis direction, when xa <xc <xb or xc <xa <xb, it is determined to be convex in the + x direction, and xb <xa <xc or xb < When xc <xa, it is determined to be convex in the −x direction. Otherwise, the edge is considered a straight line.

The shape of a pair of edges designated as described above is determined, an edge profile is calculated, a moving average profile indicating edge characteristics is created, and a uniquely determined edge position is detected.

As described above, when measuring the distance between edges and the edge length is short like the tip of the line pattern, first specify the tip as the measurement target area, An edge profile is created based on the position coordinates of the edge shape of the pattern in the region. A flat portion is detected from the edge profile, and the average position is calculated as the edge position. As one method for detecting the flat portion, the edge profile is differentiated, and the position where the differential value is first zero is obtained as the start position of the flat portion and the position where the differential value is finally zero as the end position. Can do. In addition, when a flat portion cannot be detected because there is no point where the differential value becomes zero, the profile of the flat portion estimated from the design value of the pattern width is moved and averaged to obtain an edge characteristic curve (moving average profile). And the peak position is calculated as the edge center position. The center edge position is calculated for each pattern of the pair of designated measurement target areas, and the distance between the center edges is defined as the distance between the edges. Thereby, it is possible to prevent the position of the edge from being different due to the shift of the measurement target region, and it is possible to measure the distance between the edges with high reproducibility.

(Measurement method between pattern edges)
Next, the length measurement process between edges will be described with reference to the flowchart of FIG. In the length measurement process between the edges, it is assumed that two edge positions to be measured are designated by a pair of ROI boxes (measurement target areas). In the specification by the ROI box, it is assumed that the size includes the target edge and the edge is included even if the ROI box is shifted.

First, in step S11, an SEM image including the measurement target region designated by the designated ROI box is acquired. The SEM image data is extracted from the storage unit 55 stored as pixel data.

In the next step S12, the edge shape (edge profile) of the pattern in the measurement target region is detected. This edge shape is represented by position coordinates.

The processing for detecting the edge shape of the pattern in the measurement target area will be described with reference to FIGS. FIG. 9 is a flowchart showing edge profile detection processing in the measurement target region. FIG. 10 is a diagram for explaining edge profile detection in the measurement target region. In this edge profile detection process, the edge of the pattern in the measurement target region is targeted, and the start position of the edge profile detection is the edge where the ROI box and the pattern intersect.

First, initial setting is performed in step S21 of FIG. In the initial setting, a predetermined interval (hereinafter referred to as a designation step) for detecting an edge within the ROI box designation range is designated. For example, this designation step is a distance corresponding to a predetermined number of pixels. A counter k indicating the position of the detected edge within the ROI box designation range is set to 0.

In the next step S22 to step S24, an edge position at a position away from the start position ES by a predetermined designated step d is detected.

In step S22, a temporary edge is detected at a position separated from the start position ES by a distance of (designated step d × 2). Specifically, as shown in FIG. 10 (a), a line HL orthogonal to the straight line VL from the start position ES to the lower side (−Y direction) of FIG. 10 (a) at the position (designated step d × 2). and a reference line of the profile creation, detects the edge E ₁₁ by creating a line profile. The detected edge E ₁₁ as the temporary detection edge E _11.

In the next step S23, it performs a redetection provisional detection edge E ₁₁ detected in step S22. From the start position ES and the tentative detected edge position E ₁₁ and the straight line on the starting position ES connecting the line perpendicular at a position apart a distance (specified step d × 2) and the reference line of the profile creation, the reference line A line profile is obtained, and the temporarily detected edge position is detected again. By re-detecting the temporarily detected edge position, the distance from the start position ES is made closer to (designated step d × 2).

In the next step S24, the first edge position is detected. A line profile is determined on a line orthogonal to the straight line IL ₁ connecting the start position ES and the re-detected temporary detection edge position E ₁₂ and the intermediate position MP ₁ , and an edge EP _k (x _k , y _k ) is detected. . In FIG. 10B, the edge EP ₁ is detected as the _first edge.

In the next step S25, the edge EP _k (x _k , y _k ) is set as a starting point for the next edge detection. In FIG. 10C, the edge EP ₁ is the starting point.

In the next step S26 to step S28, an edge position EP _{k + 1} (x _{k + 1} , y _{k + 1} ) that is a specified step away from the starting edge position EP _k (x _k , y _k ) is detected.

In step S26, the reference profile creating a line perpendicular at a distance and provisional detection edge E _12, which is re-detected starting point EP ₁ from the starting point EP ₁ on the straight line IL ₂ that connects only (designated step d × 2) As a line, a line profile is created to detect an edge. The detected edge as the temporary detection edge E _21.

In the next step S27, similarly to step S24, straight line of connecting the starting point EP ₁ and the tentative detected edge position E _21, perpendicular in a position apart a distance from the origin EP ₁ (specified step d × 2) The line is used as a reference line for profile creation, a line profile on this reference line is obtained, and the temporarily detected edge position is redetected.

In the next step S28, the straight line IL ₃ connecting the tentative detection edge position E _22, which is re-detected the starting point EP _1, obtains a line profile on the line perpendicular at the intermediate position MP _2, detects the edge EP k _{+ 1} To do. In FIG. 10D, the edge EP ₂ is detected as the _second edge.

In the next step S29, it is determined whether or not all edges around the pattern have been detected. If it is determined that all have been detected, the process ends. If it is determined that the detection has not yet been completed, the process proceeds to step S30.

In the next step S30, k = k + 1 is set, and the process proceeds to step S25 to detect the next edge position.

The edge position in the measurement target area is detected by the above edge shape detection process, and an edge profile is created.

Next, in step S13 in FIG. 8, a moving average width for performing a moving average on the edge profile created in step S12 is calculated. The moving average width is the length of the flat portion of the edge of the pattern, and is obtained from the point at which the differential value becomes zero first and the last zero point by differentiating the edge profile. When the point at which the differential value becomes zero is not detected, the length is obtained by subtracting a predetermined value, for example, 100 nm from the design value of the pattern width. The reason why such a predetermined value is used is that it is assumed that the corners of the actually formed pattern are not formed at right angles but the flat portion is shortened.

In the next step S14, a moving average process is performed on the edge profile detected in step S12 to calculate a moving average profile. This moving average process is calculated by moving and averaging the edge profile with the moving average width calculated in step S13. That is, an average value is calculated for each measurement point of the edge profile by the number of measurement points corresponding to the moving average width, and a moving average profile is created.

In the next step S15, the edge position (center edge position) in the measurement target area is detected based on the moving average profile. If the moving average width is the same as the width of the flat portion of the edge of the actual pattern, the peak value of the moving average profile indicates the center position of the flat portion of the edge within the range specified by the ROI. The position is set as the center edge position.

In the next step S16, the distance between the edges is calculated using the center edge position calculated for the edge of each measurement target region.

Hereinafter, calculation of the distance between edges will be described with reference to FIG. As the distance between edges, (a) the calculation of the distance in the y-axis direction between two patterns, (b) the calculation of the distance in the x-axis direction between two patterns, and (c) the shortest distance between the two patterns The calculation will be described.

(A) Calculation of distance in y-axis direction between two patterns In order to calculate the distance in the y-axis direction between patterns shown in FIG. 3 (a), a pair of ROI boxes (ROI11 and ROI12) are shown in FIG. ) Is set. Through the processing up to step S16, the position coordinate of the peak value of the moving average profile of the edge E12 in the ROI 12 is calculated as C12 (x12, y12), and the position coordinate of the peak value of the moving average profile of the edge E11 in the ROI 11 is C11. It is assumed that (x11, y11) is calculated. In this case, the y coordinate y11 of C11 is the minimum value of the edge E11, and the y coordinate y12 of C12 is the maximum value of the edge E12. Accordingly, the distance d1 in the y direction between the two edges is calculated as y11-y12.

(B) Calculation of distance in the x-axis direction between two patterns In order to calculate the distance in the x-axis direction between the patterns shown in FIG. 3 (c), a pair of ROI boxes (ROI 31 and ROI 32) are shown in FIG. ) Is set. By the processing up to step S16, the position coordinate of the peak value of the moving average profile of the edge E31 in the ROI 31 is C31 (x31, y31), and the position coordinate of the peak value of the moving average profile of the edge E32 in the ROI 32 is C32 (x32). , Y32). In this case, the x coordinate x31 of C31 is the maximum value of the edge E31, and the x coordinate x32 of C32 is the minimum value of the edge E32. Therefore, the distance d4 in the x direction between the two edges is calculated as x32-x31.

(C) Calculation of the shortest distance between two patterns Calculation of the shortest distance between edges will be described below. FIG. 3B shows two patterns (P21 and P22) formed adjacent to each other in the y-axis direction and shifted in the x-axis direction. A distance d2 in the y-axis direction between the two patterns P21 and P22 is calculated by the method (a).

Here, calculation of the shortest distance between patterns will be described. First, the edge profile of the pattern in the pair of measurement target regions (ROI21, ROI22) is smoothed by the number of predetermined measurement points. This smoothing process is performed by averaging the position coordinates of the edge profile with, for example, the position coordinates of three adjacent points. As a result, the position coordinates (smoothing profile) of the edge shape with reduced noise are calculated.

Distance √ ((x _j −x _i ) ² + (y _j −y _i ) between the measurement point (x _i , y _i ) in the ROI 21 after the smoothing processing and the measurement point (x _j , y _j ) in the ROI 22 ² ) is calculated between all measurement points, and the minimum value and position coordinates are obtained. For example, if the minimum value is taken at the edge position N21 (x21, y21) in the ROI 21 and the edge position N22 (x22, y22) in the ROI 22, the shortest distance d3 between the two edges is √ (((x21−x22 ) ² + (y21−y22) ² )

As described above, in the pattern length measurement method of this embodiment, in order to specify the position of the edge of the pattern specified by the ROI, the length of the flat part of the edge is indicated by the position coordinate as the moving average width. The edge characteristics are calculated by moving and averaging the edge profiles representing the shape of the edge. This edge characteristic uniquely indicates the center position of the flat portion of the edge. Then, the distance between the edges of the pattern designated by the pair of ROIs is calculated using the center position. This makes it possible to calculate the distance between edges with good reproducibility.

In this embodiment, the detection of the distance between the edges of the pattern has been described for the case where the pattern is formed on a substrate such as glass. However, the present invention is not limited to this, and the substrate is formed by cutting the substrate such as glass. It is also applicable to the pattern to be applied.

In the present embodiment, the detection of the distance between the edges of the two patterns has been described. However, the present invention is not limited to this, and can be applied to the detection of the distance between the edges of one pattern.

Claims

An electron beam irradiation unit for irradiating the sample while scanning the electron beam;
An image data acquisition unit that acquires an image of the pattern based on the amount of electrons generated from the sample on which the pattern is formed by irradiation of the electron beam;
A measurement target area setting unit that sets a pair of measurement areas including an edge of the image of the pattern;
A controller that detects an edge shape of the pattern in the measurement region and calculates a distance between the edges of the pattern in the pair of measurement regions;
The control unit detects a flat portion of an edge profile in which an edge of a pattern in the measurement region is indicated by position coordinates of measurement points at a predetermined interval, and an average position of the flat portion is detected in the pattern in the measurement region. The pattern length measuring device is characterized by the edge position of
When the flat portion is not detected, the control unit calculates the edge by moving and averaging the edge profile by a value obtained by subtracting a predetermined value assumed to be curved when a pattern is formed from a design value of the pattern width. The pattern length measuring apparatus according to claim 1, wherein a characteristic curve is calculated, and a position of a peak value of the edge characteristic curve is set as an edge position of the pattern.
The control unit is configured such that, of the pair of measurement areas facing in the y-axis direction, the y coordinate value of the edge position of the pattern in one measurement area is y1, and the edge position of the pattern in the other measurement area is The pattern length measuring apparatus according to claim 1, wherein when the value of the y coordinate is y2 and y1> y2, y1-y2 is a distance between edges in the y-axis direction.
The control unit has an x coordinate value of the edge position of the pattern in one measurement region of the pair of measurement regions facing in the x-axis direction, and the edge position of the pattern in the other measurement region. 2. The pattern length measuring apparatus according to claim 1, wherein when x-coordinate value is x2 and x1 <x2, x2-x1 is a distance between edges in the x-axis direction.
The control unit uses the position coordinates of each measurement point of the edge subjected to smoothing processing with a predetermined number of measurement points to reduce noise as necessary for the edge profile, and in one measurement region. The pattern length measuring apparatus according to claim 1, wherein the shortest distance between the edge of the pattern and the edge of the pattern in the other measurement region is detected.
When the edge has a flat portion and the flat portion is parallel to the x-axis, the control unit performs a moving average of the edge profile with a moving average width longer than the flat portion of the edge, and the center of the edge Calculating an x-coordinate value of the position, calculating an average value of y-coordinate values of the edge profile in a predetermined range with respect to the x-coordinate value, and calculating an edge position of the pattern in the measurement region; The pattern length measuring apparatus according to claim 1, wherein
An image for acquiring an image of the pattern based on the amount of electrons generated from the sample on which the pattern is formed by the electron beam irradiation, and an electron beam irradiation unit that irradiates the sample while scanning the sample with the electron beam In a pattern length measuring apparatus having a data acquisition unit and a measurement target region setting unit that sets a pair of measurement regions including an edge of the pattern image,
Acquiring a pattern image of the measurement target region;
Detecting an edge profile in which the edge of the pattern in the measurement target region is indicated by the position coordinates of the measurement points at a predetermined interval;
Detecting a flat portion of the edge profile;
Calculating an average position of the flat portion and determining an edge position of the pattern;
Detecting a distance between edges of the pair of patterns in the pair of measurement target regions.
When the flat portion is not detected, the step of detecting the flat portion and the step of determining the edge position include:
Calculating an edge characteristic curve by moving and averaging the edge profile with a value obtained by subtracting a predetermined value assumed to be curved when a pattern is formed from a design value of the pattern width; and The pattern length measuring method according to claim 7, wherein the pattern value measuring step is a step of determining a position of a peak value as an edge position of the pattern.
Detecting the distance between the edges comprises:
Of the pair of measurement areas facing in the y-axis direction, the y-coordinate value of the edge position of the pattern in one measurement area is y1, and the y-coordinate value of the edge position of the pattern in the other measurement area is 8. The pattern length measuring method according to claim 7, wherein y2 is a step of setting y1-y2 as a distance between edges in the y-axis direction when y1> y2.
Detecting the distance between the edges comprises:
Of the pair of measurement areas facing in the x-axis direction, the x coordinate value of the edge position of the pattern in one measurement area is x1, and the x coordinate value of the edge position of the pattern in the other measurement area is 8. The pattern length measuring method according to claim 7, wherein x2 is a step of setting x2-x1 as a distance between edges in the x-axis direction when x1 <x2.
Detecting the distance between the edges comprises:
Smoothing the edge profile by averaging the predetermined number of measurement points;
Using the position coordinates of each measurement point of the smoothed edge of the pattern in the pair of measurement areas, the shortest distance between the edge of the pattern in one measurement area and the edge of the pattern in the other measurement area is determined. The pattern length measuring method according to claim 7, further comprising a detecting step.
The step of calculating the edge characteristic curve and the step of specifying the edge position include:
When the edge has a flat portion and the flat portion is parallel to the x-axis, the edge profile is moved and averaged with a moving average width longer than the flat portion of the edge, and the x coordinate value of the center position of the edge Calculating steps,
9. The step of calculating an average value of y coordinate values of the edge profile in a predetermined range with respect to the x coordinate value, and calculating an edge position of the pattern in the measurement region. Pattern length measuring method described in 1.